Mailing-List: contact dev-help@ambari.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@ambari.apache.org
Content-Type: multipart/alternative;
 boundary="===============4441945279412824133=="
MIME-Version: 1.0
Subject: Re: Review Request 28519: Components should indicate Security State
 (via ambari-agent)
From: "Jayush Luniya" <jluniya@hortonworks.com>
To: "Robert Nettleton" <rnettleton@hortonworks.com>,
 "Jayush Luniya" <jluniya@hortonworks.com>,
 "Jaimin Jetly" <jaimin@hortonworks.com>,
 "Jonathan Hurley" <jhurley@hortonworks.com>,
 "Nate Cole" <ncole@hortonworks.com>,
 "Tom Beerbower" <tbeerbower@hortonworks.com>
Cc: "Ambari" <dev@ambari.apache.org>, "Robert Levas" <rlevas@hortonworks.com>
Date: Fri, 12 Dec 2014 05:53:46 -0000
Message-ID: <20141212055346.26419.33906@reviews.apache.org>
Auto-Submitted: auto-generated
Sender: "Jayush Luniya" <noreply@reviews.apache.org>
References: <20141210121112.5765.23409@reviews.apache.org>
In-Reply-To: <20141210121112.5765.23409@reviews.apache.org>
Reply-To: "Jayush Luniya" <jluniya@hortonworks.com>

--===============4441945279412824133==
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/28519/#review64871
-----------------------------------------------------------

Ship it!


Ship It!

- Jayush Luniya


On Dec. 10, 2014, 12:11 p.m., Robert Levas wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/28519/
> -----------------------------------------------------------
> 
> (Updated Dec. 10, 2014, 12:11 p.m.)
> 
> 
> Review request for Ambari, Jaimin Jetly, Jonathan Hurley, Jayush Luniya, Nate Cole, Robert Nettleton, and Tom Beerbower.
> 
> 
> Bugs: AMBARI-8343
>     https://issues.apache.org/jira/browse/AMBARI-8343
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> In order to properly handle the automated installation or removal of a security infrastructure (like Kerberos) in the cluster, Ambari needs to know whether each component on the hosts of the cluster is properly _secured_ or not.  This information may be compared with data on the Ambari server to help determine what steps should be taken to ensure the cluster is in the correct _secured_ state.
> 
> To do this, the current and desired component security state is maintained in the Ambari database.  The Ambari server will update the desired state details according to whether the cluster is to be secured or not and whether the relevant service has enough metadata to be secured.  If the desired and actual security state details do not match, the Ambari server will take the necessary steps to work towards synchronization. 
> 
> In order for a component to indicate its security status, a new property needs to be returned in the `STATUS_COMMAND` response message (from the Ambari agent).  This property should be named ‘securityState’ and should have one of the following values:
> 
> * `UNKNOWN` - Indicates that it is not known whether the service or component is secured or not
> * `UNSECURED` - Indicates service or component is not or should not be secured
> * `SECURED_KERBEROS` - Indicates component is or should be secured using Kerberos
> * `ERROR` - Indicates the component is not secured due to an error condition
> 
> To properly set this state value, a call needs to be executed per component querying for its specific state.  Due to the differences on how each component is secured and how it may be determined if security is setup what type is configured, and working is it properly, it is necessary for each component to have its own logic for determining this state. Therefore the ambari-agent process will need to call into the component’s configured (lifecycle) script and wait for its response - not unlike how it determines whether the component is up and running.
> 
> After the infrastructure is in place, each service definition needs to be updated to implement the new security status check function.  The function should perform the following steps:
> 
> * Determine if security is enabled or disabled
> ** If disabled, return "UNSECURED"
> ** If enabled, determine what type of security is enabled
> *** If Kerberos is configured
> **** Perform tests (kinit?, ping KDC?) to determine if the configuration appears to be working
> ***** If working, return “SECURED_KERBEROS”
> ***** If not working, return “ERROR”
> *** Else, return "UNKNOWN"
> 
> If no function is available, the Ambari agent should return “UNKNOWN”.
> 
> 
> Diffs
> -----
> 
>   ambari-agent/src/main/python/ambari_agent/ActionQueue.py fbde26f 
>   ambari-agent/src/main/python/ambari_agent/CustomServiceOrchestrator.py 08dddae 
>   ambari-agent/src/test/python/ambari_agent/TestActionQueue.py 52b586b 
>   ambari-agent/src/test/python/ambari_agent/TestCustomServiceOrchestrator.py 24ee259 
> 
> Diff: https://reviews.apache.org/r/28519/diff/
> 
> 
> Testing
> -------
> 
> Updated the following unit tests:
> * ambari-agent/src/test/python/ambari_agent/TestCustomServiceOrchestrator.py
> * ambari-agent/src/test/python/ambari_agent/TestActionQueue.py
> 
> 
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 03:06 min
> [INFO] Finished at: 2014-11-27T15:35:18+00:00
> [INFO] Final Memory: 80M/411M
> [INFO] ------------------------------------------------------------------------
> 
> 
> Thanks,
> 
> Robert Levas
> 
>


--===============4441945279412824133==--